man.c

   1 /*      $Id: man.c,v 1.87 2010/08/20 01:02:07 schwarze Exp $ */
   2 /*
   3  * Copyright (c) 2008, 2009, 2010 Kristaps Dzonsons <kristaps@bsd.lv>
   4  *
   5  * Permission to use, copy, modify, and distribute this software for any
   6  * purpose with or without fee is hereby granted, provided that the above
   7  * copyright notice and this permission notice appear in all copies.
   8  *
   9  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
  10  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
  11  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
  12  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
  13  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
  14  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
  15  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
  16  */
  17 #ifdef HAVE_CONFIG_H
  18 #include "config.h"
  19 #endif
  20
  21 #include <sys/types.h>
  22
  23 #include <assert.h>
  24 #include <stdarg.h>
  25 #include <stdlib.h>
  26 #include <stdio.h>
  27 #include <string.h>
  28
  29 #include "mandoc.h"
  30 #include "libman.h"
  31 #include "libmandoc.h"
  32
  33 const   char *const __man_macronames[MAN_MAX] = {
  34         "br",           "TH",           "SH",           "SS",
  35         "TP",           "LP",           "PP",           "P",
  36         "IP",           "HP",           "SM",           "SB",
  37         "BI",           "IB",           "BR",           "RB",
  38         "R",            "B",            "I",            "IR",
  39         "RI",           "na",           "i",            "sp",
  40         "nf",           "fi",           "r",            "RE",
  41         "RS",           "DT",           "UC",           "PD",
  42         "Sp",           "Vb",           "Ve",           "AT",
  43         "in"
  44         };
  45
  46 const   char * const *man_macronames = __man_macronames;
  47
  48 static  struct man_node *man_node_alloc(int, int,
  49                                 enum man_type, enum mant);
  50 static  int              man_node_append(struct man *,
  51                                 struct man_node *);
  52 static  void             man_node_free(struct man_node *);
  53 static  void             man_node_unlink(struct man *,
  54                                 struct man_node *);
  55 static  int              man_ptext(struct man *, int, char *, int);
  56 static  int              man_pmacro(struct man *, int, char *, int);
  57 static  void             man_free1(struct man *);
  58 static  void             man_alloc1(struct man *);
  59
  60
  61 const struct man_node *
  62 man_node(const struct man *m)
  63 {
  64
  65         return(MAN_HALT & m->flags ? NULL : m->first);
  66 }
  67
  68
  69 const struct man_meta *
  70 man_meta(const struct man *m)
  71 {
  72
  73         return(MAN_HALT & m->flags ? NULL : &m->meta);
  74 }
  75
  76
  77 void
  78 man_reset(struct man *man)
  79 {
  80
  81         man_free1(man);
  82         man_alloc1(man);
  83 }
  84
  85
  86 void
  87 man_free(struct man *man)
  88 {
  89
  90         man_free1(man);
  91         free(man);
  92 }
  93
  94
  95 struct man *
  96 man_alloc(struct regset *regs, void *data, mandocmsg msg)
  97 {
  98         struct man      *p;
  99
 100         p = mandoc_calloc(1, sizeof(struct man));
 101
 102         man_hash_init();
 103         p->data = data;
 104         p->msg = msg;
 105         p->regs = regs;
 106
 107         man_alloc1(p);
 108         return(p);
 109 }
 110
 111
 112 int
 113 man_endparse(struct man *m)
 114 {
 115
 116         if (MAN_HALT & m->flags)
 117                 return(0);
 118         else if (man_macroend(m))
 119                 return(1);
 120         m->flags |= MAN_HALT;
 121         return(0);
 122 }
 123
 124
 125 int
 126 man_parseln(struct man *m, int ln, char *buf, int offs)
 127 {
 128
 129         if (MAN_HALT & m->flags)
 130                 return(0);
 131
 132         return(('.' == buf[offs] || '\'' == buf[offs]) ?
 133                         man_pmacro(m, ln, buf, offs) :
 134                         man_ptext(m, ln, buf, offs));
 135 }
 136
 137
 138 static void
 139 man_free1(struct man *man)
 140 {
 141
 142         if (man->first)
 143                 man_node_delete(man, man->first);
 144         if (man->meta.title)
 145                 free(man->meta.title);
 146         if (man->meta.source)
 147                 free(man->meta.source);
 148         if (man->meta.rawdate)
 149                 free(man->meta.rawdate);
 150         if (man->meta.vol)
 151                 free(man->meta.vol);
 152         if (man->meta.msec)
 153                 free(man->meta.msec);
 154 }
 155
 156
 157 static void
 158 man_alloc1(struct man *m)
 159 {
 160
 161         memset(&m->meta, 0, sizeof(struct man_meta));
 162         m->flags = 0;
 163         m->last = mandoc_calloc(1, sizeof(struct man_node));
 164         m->first = m->last;
 165         m->last->type = MAN_ROOT;
 166         m->last->tok = MAN_MAX;
 167         m->next = MAN_NEXT_CHILD;
 168 }
 169
 170
 171 static int
 172 man_node_append(struct man *man, struct man_node *p)
 173 {
 174
 175         assert(man->last);
 176         assert(man->first);
 177         assert(MAN_ROOT != p->type);
 178
 179         switch (man->next) {
 180         case (MAN_NEXT_SIBLING):
 181                 man->last->next = p;
 182                 p->prev = man->last;
 183                 p->parent = man->last->parent;
 184                 break;
 185         case (MAN_NEXT_CHILD):
 186                 man->last->child = p;
 187                 p->parent = man->last;
 188                 break;
 189         default:
 190                 abort();
 191                 /* NOTREACHED */
 192         }
 193
 194         assert(p->parent);
 195         p->parent->nchild++;
 196
 197         if ( ! man_valid_pre(man, p))
 198                 return(0);
 199
 200         switch (p->type) {
 201         case (MAN_HEAD):
 202                 assert(MAN_BLOCK == p->parent->type);
 203                 p->parent->head = p;
 204                 break;
 205         case (MAN_BODY):
 206                 assert(MAN_BLOCK == p->parent->type);
 207                 p->parent->body = p;
 208                 break;
 209         default:
 210                 break;
 211         }
 212
 213         man->last = p;
 214
 215         switch (p->type) {
 216         case (MAN_TEXT):
 217                 if ( ! man_valid_post(man))
 218                         return(0);
 219                 if ( ! man_action_post(man))
 220                         return(0);
 221                 break;
 222         default:
 223                 break;
 224         }
 225
 226         return(1);
 227 }
 228
 229
 230 static struct man_node *
 231 man_node_alloc(int line, int pos, enum man_type type, enum mant tok)
 232 {
 233         struct man_node *p;
 234
 235         p = mandoc_calloc(1, sizeof(struct man_node));
 236         p->line = line;
 237         p->pos = pos;
 238         p->type = type;
 239         p->tok = tok;
 240         return(p);
 241 }
 242
 243
 244 int
 245 man_elem_alloc(struct man *m, int line, int pos, enum mant tok)
 246 {
 247         struct man_node *p;
 248
 249         p = man_node_alloc(line, pos, MAN_ELEM, tok);
 250         if ( ! man_node_append(m, p))
 251                 return(0);
 252         m->next = MAN_NEXT_CHILD;
 253         return(1);
 254 }
 255
 256
 257 int
 258 man_head_alloc(struct man *m, int line, int pos, enum mant tok)
 259 {
 260         struct man_node *p;
 261
 262         p = man_node_alloc(line, pos, MAN_HEAD, tok);
 263         if ( ! man_node_append(m, p))
 264                 return(0);
 265         m->next = MAN_NEXT_CHILD;
 266         return(1);
 267 }
 268
 269
 270 int
 271 man_body_alloc(struct man *m, int line, int pos, enum mant tok)
 272 {
 273         struct man_node *p;
 274
 275         p = man_node_alloc(line, pos, MAN_BODY, tok);
 276         if ( ! man_node_append(m, p))
 277                 return(0);
 278         m->next = MAN_NEXT_CHILD;
 279         return(1);
 280 }
 281
 282
 283 int
 284 man_block_alloc(struct man *m, int line, int pos, enum mant tok)
 285 {
 286         struct man_node *p;
 287
 288         p = man_node_alloc(line, pos, MAN_BLOCK, tok);
 289         if ( ! man_node_append(m, p))
 290                 return(0);
 291         m->next = MAN_NEXT_CHILD;
 292         return(1);
 293 }
 294
 295
 296 int
 297 man_word_alloc(struct man *m, int line, int pos, const char *word)
 298 {
 299         struct man_node *n;
 300         size_t           sv, len;
 301
 302         len = strlen(word);
 303
 304         n = man_node_alloc(line, pos, MAN_TEXT, MAN_MAX);
 305         n->string = mandoc_malloc(len + 1);
 306         sv = strlcpy(n->string, word, len + 1);
 307
 308         /* Prohibit truncation. */
 309         assert(sv < len + 1);
 310
 311         if ( ! man_node_append(m, n))
 312                 return(0);
 313
 314         m->next = MAN_NEXT_SIBLING;
 315         return(1);
 316 }
 317
 318
 319 /*
 320  * Free all of the resources held by a node.  This does NOT unlink a
 321  * node from its context; for that, see man_node_unlink().
 322  */
 323 static void
 324 man_node_free(struct man_node *p)
 325 {
 326
 327         if (p->string)
 328                 free(p->string);
 329         free(p);
 330 }
 331
 332
 333 void
 334 man_node_delete(struct man *m, struct man_node *p)
 335 {
 336
 337         while (p->child)
 338                 man_node_delete(m, p->child);
 339
 340         man_node_unlink(m, p);
 341         man_node_free(p);
 342 }
 343
 344
 345 static int
 346 man_ptext(struct man *m, int line, char *buf, int offs)
 347 {
 348         int              i;
 349
 350         /* Ignore bogus comments. */
 351
 352         if ('\\' == buf[offs] &&
 353                         '.' == buf[offs + 1] &&
 354                         '"' == buf[offs + 2])
 355                 return(man_pmsg(m, line, offs, MANDOCERR_BADCOMMENT));
 356
 357         /* Literal free-form text whitespace is preserved. */
 358
 359         if (MAN_LITERAL & m->flags) {
 360                 if ( ! man_word_alloc(m, line, offs, buf + offs))
 361                         return(0);
 362                 goto descope;
 363         }
 364
 365         /* Pump blank lines directly into the backend. */
 366
 367         for (i = offs; ' ' == buf[i]; i++)
 368                 /* Skip leading whitespace. */ ;
 369
 370         if ('\0' == buf[i]) {
 371                 /* Allocate a blank entry. */
 372                 if ( ! man_word_alloc(m, line, offs, ""))
 373                         return(0);
 374                 goto descope;
 375         }
 376
 377         /*
 378          * Warn if the last un-escaped character is whitespace. Then
 379          * strip away the remaining spaces (tabs stay!).
 380          */
 381
 382         i = (int)strlen(buf);
 383         assert(i);
 384
 385         if (' ' == buf[i - 1] || '\t' == buf[i - 1]) {
 386                 if (i > 1 && '\\' != buf[i - 2])
 387                         if ( ! man_pmsg(m, line, i - 1, MANDOCERR_EOLNSPACE))
 388                                 return(0);
 389
 390                 for (--i; i && ' ' == buf[i]; i--)
 391                         /* Spin back to non-space. */ ;
 392
 393                 /* Jump ahead of escaped whitespace. */
 394                 i += '\\' == buf[i] ? 2 : 1;
 395
 396                 buf[i] = '\0';
 397         }
 398
 399         if ( ! man_word_alloc(m, line, offs, buf + offs))
 400                 return(0);
 401
 402         /*
 403          * End-of-sentence check.  If the last character is an unescaped
 404          * EOS character, then flag the node as being the end of a
 405          * sentence.  The front-end will know how to interpret this.
 406          */
 407
 408         assert(i);
 409         if (mandoc_eos(buf, (size_t)i, 0))
 410                 m->last->flags |= MAN_EOS;
 411
 412 descope:
 413         /*
 414          * Co-ordinate what happens with having a next-line scope open:
 415          * first close out the element scope (if applicable), then close
 416          * out the block scope (also if applicable).
 417          */
 418
 419         if (MAN_ELINE & m->flags) {
 420                 m->flags &= ~MAN_ELINE;
 421                 if ( ! man_unscope(m, m->last->parent, MANDOCERR_MAX))
 422                         return(0);
 423         }
 424
 425         if ( ! (MAN_BLINE & m->flags))
 426                 return(1);
 427         m->flags &= ~MAN_BLINE;
 428
 429         if ( ! man_unscope(m, m->last->parent, MANDOCERR_MAX))
 430                 return(0);
 431         return(man_body_alloc(m, line, offs, m->last->tok));
 432 }
 433
 434
 435 int
 436 man_pmacro(struct man *m, int ln, char *buf, int offs)
 437 {
 438         int              i, j, ppos;
 439         enum mant        tok;
 440         char             mac[5];
 441         struct man_node *n;
 442
 443         /* Comments and empties are quickly ignored. */
 444
 445         offs++;
 446
 447         if ('\0' == buf[offs])
 448                 return(1);
 449
 450         i = offs;
 451
 452         /*
 453          * Skip whitespace between the control character and initial
 454          * text.  "Whitespace" is both spaces and tabs.
 455          */
 456
 457         if (' ' == buf[i] || '\t' == buf[i]) {
 458                 i++;
 459                 while (buf[i] && (' ' == buf[i] || '\t' == buf[i]))
 460                         i++;
 461                 if ('\0' == buf[i])
 462                         goto out;
 463         }
 464
 465         ppos = i;
 466
 467         /*
 468          * Copy the first word into a nil-terminated buffer.
 469          * Stop copying when a tab, space, or eoln is encountered.
 470          */
 471
 472         j = 0;
 473         while (j < 4 && '\0' != buf[i] && ' ' != buf[i] && '\t' != buf[i])
 474                 mac[j++] = buf[i++];
 475         mac[j] = '\0';
 476
 477         tok = (j > 0 && j < 4) ? man_hash_find(mac) : MAN_MAX;
 478         if (MAN_MAX == tok) {
 479                 man_vmsg(m, MANDOCERR_MACRO, ln, ppos,
 480                     "unknown macro: %s%s",
 481                     buf, strlen(buf) > 3 ? "..." : "");
 482                 return(1);
 483         }
 484
 485         /* The macro is sane.  Jump to the next word. */
 486
 487         while (buf[i] && ' ' == buf[i])
 488                 i++;
 489
 490         /*
 491          * Trailing whitespace.  Note that tabs are allowed to be passed
 492          * into the parser as "text", so we only warn about spaces here.
 493          */
 494
 495         if ('\0' == buf[i] && ' ' == buf[i - 1])
 496                 if ( ! man_pmsg(m, ln, i - 1, MANDOCERR_EOLNSPACE))
 497                         goto err;
 498
 499         /*
 500          * Remove prior ELINE macro, as it's being clobbering by a new
 501          * macro.  Note that NSCOPED macros do not close out ELINE
 502          * macros---they don't print text---so we let those slip by.
 503          */
 504
 505         if ( ! (MAN_NSCOPED & man_macros[tok].flags) &&
 506                         m->flags & MAN_ELINE) {
 507                 assert(MAN_TEXT != m->last->type);
 508
 509                 /*
 510                  * This occurs in the following construction:
 511                  *   .B
 512                  *   .br
 513                  *   .B
 514                  *   .br
 515                  *   I hate man macros.
 516                  * Flat-out disallow this madness.
 517                  */
 518                 if (MAN_NSCOPED & man_macros[m->last->tok].flags) {
 519                         man_pmsg(m, ln, ppos, MANDOCERR_SYNTLINESCOPE);
 520                         return(0);
 521                 }
 522
 523                 n = m->last;
 524
 525                 assert(n);
 526                 assert(NULL == n->child);
 527                 assert(0 == n->nchild);
 528
 529                 if ( ! man_nmsg(m, n, MANDOCERR_LINESCOPE))
 530                         return(0);
 531
 532                 man_node_delete(m, n);
 533                 m->flags &= ~MAN_ELINE;
 534         }
 535
 536         /*
 537          * Save the fact that we're in the next-line for a block.  In
 538          * this way, embedded roff instructions can "remember" state
 539          * when they exit.
 540          */
 541
 542         if (MAN_BLINE & m->flags)
 543                 m->flags |= MAN_BPLINE;
 544
 545         /* Call to handler... */
 546
 547         assert(man_macros[tok].fp);
 548         if ( ! (*man_macros[tok].fp)(m, tok, ln, ppos, &i, buf))
 549                 goto err;
 550
 551 out:
 552         /*
 553          * We weren't in a block-line scope when entering the
 554          * above-parsed macro, so return.
 555          */
 556
 557         if ( ! (MAN_BPLINE & m->flags)) {
 558                 m->flags &= ~MAN_ILINE;
 559                 return(1);
 560         }
 561         m->flags &= ~MAN_BPLINE;
 562
 563         /*
 564          * If we're in a block scope, then allow this macro to slip by
 565          * without closing scope around it.
 566          */
 567
 568         if (MAN_ILINE & m->flags) {
 569                 m->flags &= ~MAN_ILINE;
 570                 return(1);
 571         }
 572
 573         /*
 574          * If we've opened a new next-line element scope, then return
 575          * now, as the next line will close out the block scope.
 576          */
 577
 578         if (MAN_ELINE & m->flags)
 579                 return(1);
 580
 581         /* Close out the block scope opened in the prior line.  */
 582
 583         assert(MAN_BLINE & m->flags);
 584         m->flags &= ~MAN_BLINE;
 585
 586         if ( ! man_unscope(m, m->last->parent, MANDOCERR_MAX))
 587                 return(0);
 588         return(man_body_alloc(m, ln, offs, m->last->tok));
 589
 590 err:    /* Error out. */
 591
 592         m->flags |= MAN_HALT;
 593         return(0);
 594 }
 595
 596
 597 int
 598 man_vmsg(struct man *man, enum mandocerr t,
 599                 int ln, int pos, const char *fmt, ...)
 600 {
 601         char             buf[256];
 602         va_list          ap;
 603
 604         va_start(ap, fmt);
 605         vsnprintf(buf, sizeof(buf) - 1, fmt, ap);
 606         va_end(ap);
 607         return((*man->msg)(t, man->data, ln, pos, buf));
 608 }
 609
 610
 611 /*
 612  * Unlink a node from its context.  If "m" is provided, the last parse
 613  * point will also be adjusted accordingly.
 614  */
 615 static void
 616 man_node_unlink(struct man *m, struct man_node *n)
 617 {
 618
 619         /* Adjust siblings. */
 620
 621         if (n->prev)
 622                 n->prev->next = n->next;
 623         if (n->next)
 624                 n->next->prev = n->prev;
 625
 626         /* Adjust parent. */
 627
 628         if (n->parent) {
 629                 n->parent->nchild--;
 630                 if (n->parent->child == n)
 631                         n->parent->child = n->prev ? n->prev : n->next;
 632         }
 633
 634         /* Adjust parse point, if applicable. */
 635
 636         if (m && m->last == n) {
 637                 /*XXX: this can occur when bailing from validation. */
 638                 /*assert(NULL == n->next);*/
 639                 if (n->prev) {
 640                         m->last = n->prev;
 641                         m->next = MAN_NEXT_SIBLING;
 642                 } else {
 643                         m->last = n->parent;
 644                         m->next = MAN_NEXT_CHILD;
 645                 }
 646         }
 647
 648         if (m && m->first == n)
 649                 m->first = NULL;
 650 }